A vehicle is mounted with many sensors, including a temperature sensor that measures the temperature of the interior of an engine, a throttle position sensor that detects a throttle opening, and the like. In on-board units, various measurements are carried out by detecting the voltage levels of input signals from these sensors by AD conversion or carrying out other like processing. In these on-board units, any sensor failure has been conventionally detected based on the voltage level of an input signal detected by AD conversion. In in-vehicle applications, sensors are in such an environment that they are susceptible to external noise; therefore, it is necessary to widen the dynamic range of each sensor to minimize the influence of external noise. If an open failure occurs in an input terminal for input signals of IC that carries out AD conversion, the following will take place: the voltage level of input to the AD conversion circuit is made undefined by parasitic capacitance produced in the IC; and as a result, the open failure in the input terminal cannot be detected from the result of AD conversion. One of methods for detecting an input terminal open failure and the like is such that: multiplexing is implemented using multiple sensors, multiple input circuits, and multiple AD converters and it is determined from the relation between multiple AD conversion results. However, this inevitably leads to an increased cost.
To cope with this, Patent Document 1 describes a break detector including: an input terminal to which input signals are inputted from a sensor and the like; an AD converter connected to the input terminal through a signal line; a capacitor provided between the signal line and another line; and SW provided on the signal line for controlling the connection between the capacitor and the input terminal. In this break detector, the capacitor and the input terminal are disconnected from each other by the SW during periods for which AD conversion is not carried out; and the capacitor is connected to a negative power supply and the voltage across the capacitor is set to a negative voltage out of a variation range of the voltages of input signals. When AD conversion is carried out, the capacitor, input terminal, and AD converter are connected together by the SW. During the AD conversion, the voltage value of the input terminal is detected at normal times and a negative voltage generated by the capacitor is detected when an open failure occurs in the input terminal. In the invention described in Patent Document 1, therefore, the AD converter must be capable of detecting a wider range of voltages than the variation range of input signals and this complicates the configuration of the AD converter and increases its size.
With the break detector described in Patent Document 1, any open failure in the input terminal can be detected. However, it is necessary to apply negative voltage to the capacitor for a certain time to set a negative voltage in the capacitor before starting AD conversion. This delays detection of the voltage value of an input signal.
According to the invention described in Patent Document 1, any open failure in an input terminal and the like can be detected. However, it is necessary to detect a voltage value out of the variation range of an input signal by an AD converter. For this reason, it is necessary to adjust a sensor and the like that generate input signals to make the variation range of the input signals narrower than the range of voltage that can be detected with the AD converter. Or, it is necessary to configure the AD converter so that it can detect a wider range of voltage than the variation range of input signals. For this reason, there is a possibility that a used sensor and the like are limited or the configuration of the AD converter is complicated or the size thereof is increased.
Especially, with respect to AD converters for in-vehicle applications that are susceptible to external noise, as already mentioned above, it is necessary to widen the dynamic range of each sensor to minimize the influence of external noise. It is difficult to narrow the variation range of input signals from each sensor. For this reason, it is necessary to configure an AD converter so that it can detect a wider range of voltage than the variation range of input signals. This causes the problem of the complicated configuration and increased size of the AD converter.
Further, according to the invention described in Patent Document 1, an open failure in an input terminal and the like can be detected. However, since electric charges are charged in a capacitor connected to an input terminal to detect any open failure in the input terminal or the like, a problem arises. The voltage value of an input signal is varied and there is a possibility that an error is produced in a measurement value of the voltage of the input signal.    [Patent Document 1] JP-A-2000-55966